- From: Jean-Marc Valin <jmvalin@mozilla.com>
- Date: Thu, 02 May 2013 17:21:44 -0400
- To: Frederick Umminger <frederick.umminger@gmail.com>
- CC: "public-audio@w3.org" <public-audio@w3.org>
On 02/05/13 05:04 PM, Frederick Umminger wrote: > What I was trying to say, is that if a soft-clipper clips just a little > at 0 dB FS, it will clip a lot at 6 dB over, and a huge amount at say 24 > dB over. > No matter what your soft-clipping wave-shaper curve is to start with, at > that point the output will be hard-clipped. Oh I totally agree here. Soft-clipping is useful when the level is generally below 0 dB FS with only a few spikes in the 0-6 dB range. Beyond that you definitely need something else. However, I think the idea should be to avoid getting there in the first place (using the spec and/or a limiter). > Compressors can have artifacts, but final-output limiters are designed > to be transparent. > And it is a fact that in professional audio production across a variety > of industries, limiters rather than soft-clippers are primarily used to > deal with overs. Well, why not have a hardcoded limiter that cannot be overridden? Sounds like it would provide higher quality with no way to "cheat the system". This would prevent loudness war and "ads are so loud" issues right from the start. The compressor could still be used an an optional "effect" but the clipping issue would be avoided as part of the design itself. Cheers, Jean-Marc
Received on Thursday, 2 May 2013 21:22:13 UTC